VP8 and MPEG LA (WebM blog) [LWN.net]

VP8 and MPEG LA (WebM blog)

Posted Mar 7, 2013 19:58 UTC (Thu) by drag (guest, #31333) [Link] (25 responses)

Well that is certainly a interesting announcement. I wonder how far over they bent Google for Google to get these concessions.

So much we will probably never know. At least not anytime in the next couple decades.

> It further provides for sublicensing those VP8 techniques in one successor generation to the VP8 video codec."

They started working on "VP9" yet?

VP8 and MPEG LA (WebM blog)

Posted Mar 7, 2013 19:59 UTC (Thu) by drag (guest, #31333) [Link] (24 responses)

oh silly me. I should use the googles before posting.

http://news.cnet.com/8301-1023_3-57561111-93/googles-new-...

VP8 and MPEG LA (WebM blog)

Posted Mar 7, 2013 22:05 UTC (Thu) by dowdle (subscriber, #659) [Link]

Oh look... there's a wikipedia page for that:

https://en.wikipedia.org/wiki/VP9

VP8 and MPEG LA (WebM blog)

Posted Mar 8, 2013 0:25 UTC (Fri) by Wol (subscriber, #4433) [Link] (22 responses)

Umm... apart from the fact that that page makes no mention of any money pauid by Google, but does menttion that no licence is being offered by MPEG-LA...

What's the betting that it's MPEG-LA that have had to bend over ... ?

I wouldn't put it past Google to have gone to MPEG-LA and said "look at this pile of evidence. Do you want us to give it to the USPTO and ask them to re-examine your patents? Just sign here and there'll be no unpleasantness :-)".

Cheers,
Wol

VP8 and MPEG LA (WebM blog)

Posted Mar 8, 2013 2:34 UTC (Fri) by drag (guest, #31333) [Link] (21 responses)

> What's the betting that it's MPEG-LA that have had to bend over ... ?

Because I didn't see any mention of Google licensing patents back to MPEG-LA. I think that is usually mentioned if there is any cross-licensing. (Unless Google is now a member of MPEG-LA ?)

The most likely way Google would be able to get this would be through cross-licensing patents. They paid a LOT for Motorola's patents and maybe some of the MPEG-LA members wanted a cheap way to avoid violating some of those. But since I don't see any mention of cross-licensing then I don't know how likely this is.

> but does menttion that no licence is being offered by MPEG-LA...

Yes.

I am curious how Google accomplished this. They know that VP8 would be worthless if people had to license it from MPEG-LA. It would make it effectively the same as H.264 from a end user's legal perspective. So they did something very important and impressive to have a blanket license granted to not only their VP8 implementation, but ALL VP8 implementations whether they are derivative or not. Without this it would pretty much eliminate any point in using it over the already well established H.264 stuff; both from business and open source perspective.

> I wouldn't put it past Google to have gone to MPEG-LA and said "look at this pile of evidence. Do you want us to give it to the USPTO and ask them to re-examine your patents? Just sign here and there'll be no unpleasantness :-)".

Given the history of USPTO and the massive unlikelihood that anybody would be able to defeat the patents by making them invalid in court I think this is extremely unlikely. Could be wrong, but I don't think so. It would be nice. By why play this game? Why not release the evidence and just leave it at that? They don't have to take MPEG-LA to court to convince other people that the patents have no teeth. Once people see the evidence then they would simply not feel compelled to pay MPEG-LA for the licenses over those particular patents. You don't need to court to destroy patents...

A alternative, and more somewhat more likely scenario (although I still think is unlikely), is that this is purely a marketing stunt. There is no way to prove a negative in this sort of case (MPEG-LA is not a risk). So Google figured out that the 'Fear and Doubt' game was working against VP8 adoption. The effective licensing costs for H.264 are so low that it made the legal issue mute when dealing with other large players. It was seen by large players that complying with H.264 was costing them virtually nothing and VP8 represented a unknown legal risk. So why risk it?

So Google simply threw a few million at MPEG-LA to get the blanket license for patents that seem like they could pose a risk, even if Google's lawyers think that they may be able to work around them. This way it removes the FUD and clears the way for the adoption of VP8 and VP9. This would be expensive, but they have a huge budget for promoting their products and maybe this was seen as a cheaper way to get users out of the shadow of doubt then to try to convince them through adverts and press releases.

VP8 and MPEG LA (WebM blog)

Posted Mar 8, 2013 6:02 UTC (Fri) by eru (subscriber, #2753) [Link] (7 responses)

Feeling pessimistic, I wonder it MPEG-LA co-operated just to get good PR for little cost. At this point it is fairly clear that VP8 is not a serious competitor to H.264 any more, if it ever was. Simply because software and cameras spitting out H.264 natively are now ubiquitous.

VP8 and MPEG LA (WebM blog)

Posted Mar 8, 2013 8:19 UTC (Fri) by daniel (guest, #3181) [Link] (3 responses)

I'll go with the theory that the threat of invalidating a significant number of patents in the MPEG pool was enough to make the MPEG Racketeers, err sorry, Licensing Association cooperative.

VP8 and MPEG LA (WebM blog)

Posted Mar 8, 2013 12:12 UTC (Fri) by man_ls (guest, #15091) [Link] (2 responses)

Google license H.264 right now for Chrome and YouTube, so I guess they pay MPEG-LA a lot of money. That must give them some leverage too: "Please do us a favor and license this V8 thing of ours, or the next version of the most widely used browser and the most widely used video site will not use your codec". I am guessing there is a big shiny button at YouTube HQ with the label "convert all videos to V8".

VP8 and MPEG LA (WebM blog)

Posted Mar 8, 2013 13:12 UTC (Fri) by bawjaws (guest, #56952) [Link] (1 responses)

That button got pressed a while ago, you get VP8 by default in the HTML5 player I believe.

More interestingly, the films and TV shows that Google rents are going to start being VP8 on ChromeOS due to that fake DRM they're trying to put into web standards to stop people using the fake DRM in Flash and Silverlight.

Kind of mixed feelings about that, but, it seems to be working.

VP8 and MPEG LA (WebM blog)

Posted Mar 8, 2013 13:28 UTC (Fri) by man_ls (guest, #15091) [Link]

You are right about the big shiny button. So perhaps there is another big shiny button right besides it that says: "Delete all H.264 videos right now" so only the VP8 copy is left. Think about it: Chrome can read it, Firefox can read it, Android can read it. Perhaps iOS or IE<=8 cannot access YouTube anymore? Tough luck.

It is a bit thermonuclear: as a threat it looks great, in real life it sucks.

VP8 and MPEG LA (WebM blog)

Posted Mar 8, 2013 13:24 UTC (Fri) by bawjaws (guest, #56952) [Link] (1 responses)

The reason behind this for Google seems to be to get VP8 selected as the Mandatory-to-implement video codec for WebRTC (i.e. standards-based Skype in the browser).

That's a big deal, and particularly the ability to jump quickly to VP9 due to the lesser format lock-in for effectively disposable video makes this a good place for open codecs to start.

They've already chosen Opus on the audio side, it could well be a very big deal.

Google's announcement email to the WebRTC list:
http://www.ietf.org/mail-archive/web/rtcweb/current/msg06...

They're supposed to be making the codec decision at an IETF meeting in a week or so. Google delayed addressing the topic at the last meeting saying they had something lined up that would silence the critics. I guess the other shoe just dropped.

VP8 and MPEG LA (WebM blog)

Posted Mar 8, 2013 14:10 UTC (Fri) by bawjaws (guest, #56952) [Link]

One important point that many people seemed to have overlooked.

MPEG (not MPEG-LA) have been kicking around the idea of an royalty-free codec for a while, basing it on old MPEG tech that has fallen out of patent or on getting H.264 folk to agree to freeing up some baseline patents.

With this latest announcement about patents, Google has proposed VP8 to become the official MPEG royalty free codec.

http://www.robglidden.com/2013/03/google-mpegla-vp8-mpeg-...

VP8 and MPEG LA (WebM blog)

Posted Mar 8, 2013 18:30 UTC (Fri) by Lennie (subscriber, #49641) [Link]

That can be, but what might be more interesting is the next battle, between H.265 and VP9, maybe this time it will be technical and possible financial merits (you don't have to pay for a VP9 license).

H.265 might still have a slight head-start though.

VP8 and MPEG LA (WebM blog)

Posted Mar 8, 2013 6:22 UTC (Fri) by iabervon (subscriber, #722) [Link] (5 responses)

I think the key is the FTC's "In the Matter of Motorola Mobility LLC, a limited liability company, and Google Inc., a corporation", where Apple and Microsoft have been saying that patents essential to standards should be made pretty much irrelevant, while Google has been arguing to keep them meaningful. This is, of course, because Apple owes Google tons of money for Motorola's patents on H.264 that Apple has neglected to pay for, and Google would like to use this fact to get iOS out of the US unless Apple actually bothers to pay. Meanwhile, the MPEG LA's business model relies on nobody being able to do what Apple is trying to do, and Google losing would pretty much mean that the MPEG LA has no way to make anyone pay for H.264. So Google calls up the MPEG LA and says, "We'll do the best we can, but our legal team's been getting distracted by the VP8 patent pool issue..." Between the options of people using VP8 as an alternative to paying for H.264 licenses and people using H.264 and just not paying for the licenses, the MPEG LA obviously prefers the former.

VP8 and MPEG LA (WebM blog)

Posted Mar 8, 2013 13:06 UTC (Fri) by pboddie (guest, #50784) [Link] (4 responses)

Where have Apple and Microsoft claimed that "patents essential to standards should be made pretty much irrelevant"? That would be a massive U-turn, especially for Microsoft (and some of their partners, although they never act in the interests of their partners, so that probably isn't a concern) and would probably undermine their own licensing business.

VP8 and MPEG LA (WebM blog)

Posted Mar 8, 2013 18:41 UTC (Fri) by iabervon (subscriber, #722) [Link] (3 responses)

In their comments on the FTC proposal I mentioned (at http://ftc.gov/os/comments/motorolagoogle/index.shtm). The proposal would prohibit getting injunctions based on standards-essential patents, which leaves you with nothing you can do if someone doesn't pay. I suspect that Microsoft's patents are mostly not essential (you could theoretically do it differently, since the standard relies on the outcome, not the method) or not for official standards (only needed if you want to interoperate with Microsoft products).

VP8 and MPEG LA (WebM blog)

Posted Mar 8, 2013 22:25 UTC (Fri) by pboddie (guest, #50784) [Link] (2 responses)

Skimming those submissions (because I don't have the time or inclination to read them closely), it appears that Apple and Microsoft favour FRAND-encumbered standards where the licensing terms should be decided through some process outside the courtroom. That's quite different from regarding patents in standards as irrelevant: anyone not getting a licence for some patent that is supposed to be essential to a standard is still going to get a letter asking them to pay up, but I imagine that the responsibility for deciding the matter, setting the price, and sending the letter is likely to fall on the appointed cartel rather than on some legal functionary or the lawyer of any particular company.

VP8 and MPEG LA (WebM blog)

Posted Mar 8, 2013 23:21 UTC (Fri) by iabervon (subscriber, #722) [Link] (1 responses)

Sure, they get a letter asking them to pay up. But they just ignore that letter, because there aren't any meaningful consequences to not paying, because the patent holder can't get an injunction against them. Apple doesn't mention this in their comments, but that would be an implication of the proposed FTC rule, and the whole thing came up because Apple has been ignoring that letter and Motorola wants something done about it.

Injunctions are not a replacement for damages

Posted Mar 9, 2013 3:04 UTC (Sat) by tialaramex (subscriber, #21167) [Link]

Not being able to impose an injunction doesn't mean that there are no consequences.

Injunctions are a means of _preventing harm_. The court must be persuaded that without the injunction some type of harm will occur to the plaintiff (e.g. their priceless 10th century wall painting will be destroyed, or their stranglehold on the desktop PC market will be broken) which is unlawful, and which cannot be made right through any of the court's normal remedies (e.g. monetary damages). If this is so it can decide based on very limited evidence (often timely action is essential) that the plaintiff is on balance likely to succeed in their claim and that it's not excessive to prevent the defendant from doing something (e.g. demolishing the wall, or shipping a new OS) meanwhile.

But even without an injunction all the ordinary remedies are still available. You just have to actually go through the entire lawsuit process to reach a judgement and get the remedy you want, rather than cutting straight to the injunction and hoping it strangles the defendant so badly that they've no choice but to forfeit their day in court altogether.

A world where Apple can ship infringing iPhones despite a lawsuit, but then has the bailiffs at One Infinite Loop repossessing the office furniture is not one where there "aren't any meaningful consequences" to not paying.

VP8 and MPEG LA (WebM blog)

Posted Mar 8, 2013 7:53 UTC (Fri) by epa (subscriber, #39769) [Link] (5 responses)

Google could release evidence showing that certain MPEG-LA patents are almost certain to be invalidated if tested in court. But MPEG-LA would never acknowledge this; they'd still mumble things in public about forming a patent pool and so on. Not everyone has a sharp enough in-house legal team to weigh up the true threat; most would look at the situation, see that there is some legal uncertainty, and decide to play it safe. Remember what the U in FUD stands for. The way they resolved it is much better.

VP8 and MPEG LA (WebM blog)

Posted Mar 8, 2013 9:03 UTC (Fri) by Wol (subscriber, #4433) [Link] (4 responses)

What some people seem to have missed is ...

Patent APPLICATIONS are pretty much rubber-stamped by the USPTO.

Patent RE-EXAMS are much stricter - a significant number of patents are invalidated by a re-exam.

And it wouldn't cost Google much to swamp the USPTO with re-exam requests that are on-point, pertinent, and capable of doing much much much damage to the patents.

It is NOT normal practice to invalidate patents in court. Even in a court case it is far more effective, and cheaper, to refer the patents back to the USPTO for re-exam. Only if that route fails do you really want to get into a court battle over validity.

Cheers,
Wol

VP8 and MPEG LA (WebM blog)

Posted Mar 8, 2013 9:26 UTC (Fri) by epa (subscriber, #39769) [Link]

I thought that sending patents back for re-examination was best avoided because if the USPTO re-examines it and finds it still valid, that counts as evidence in any subsequent court case.

VP8 and MPEG LA (WebM blog)

Posted Mar 8, 2013 13:16 UTC (Fri) by bawjaws (guest, #56952) [Link] (2 responses)

The MPEG-LA patents are usually very specific to what's in their standards and therefore (comparatively) solid patents that would be hard to invalidate.

Where it gets muddy is when they claim that their stack of patents apply to other codecs. That seems to involve courthouses and juries learning about the ins-and-outs of video compression, therefore high cost and risk.

Why would it be hard?

Posted Mar 8, 2013 16:05 UTC (Fri) by Wol (subscriber, #4433) [Link] (1 responses)

Just because the patent is rock-solid tied to the standard, doesn't mean there isn't a load of prior art out there.

One only has to look at the gif patent (I think I've got the right one) where the USPTO granted two different patents on THE SAME algorithm. All Google has to do (and the tighter the patent is, the easier this is to do) is find a prior implementation of the same technique, and BOOM, the patent is gone.

I can't see it being hard for Google to find prior art. After all, they are the masters of search ... :-)

Cheers,
Wol

Why would it be hard?

Posted Mar 8, 2013 21:08 UTC (Fri) by dlang (guest, #313) [Link]

land war with russia

fight with google over something involving searching....

VP8 and MPEG LA (WebM blog)

Posted Mar 9, 2013 21:38 UTC (Sat) by tterribe (guest, #66972) [Link]

> Because I didn't see any mention of Google licensing patents back to
> MPEG-LA. I think that is usually mentioned if there is any cross-
> licensing. (Unless Google is now a member of MPEG-LA ?)

Google is a licensee of the H.264 pool at least (#384 on the list at <http://www.mpegla.com/main/programs/AVC/Pages/Licensees.aspx>). That means they and their affiliates have to offer a reciprocal license to all other licensees.

This is one of the major controversies of the Motorola v. Microsoft case: since Motorola was acquired after Google agreed to the license and not included in the explicit list of "affiliates" Google provided, and since the lawsuit began before Google acquired them, Motorola claims the reciprocal license does not apply. Microsoft claims it does.

Thank you, Google!

Posted Mar 7, 2013 23:09 UTC (Thu) by boog (subscriber, #30882) [Link] (11 responses)

Brilliant news. There is hope for a free video future.

Thank you, Google!

Posted Mar 7, 2013 23:59 UTC (Thu) by Zizzle (guest, #67739) [Link] (10 responses)

I'm not so optimistic. I'm sure Apple and Google will find new reasons not to inter-operate on WebRTC and video in general.

Thank you, Google!

Posted Mar 8, 2013 2:36 UTC (Fri) by drag (guest, #31333) [Link] (9 responses)

Especially since Apple, being MPEG-LA member, has a financial and competitive advantage to people adopting MPEG codecs over free ones.

Thank you, Google!

Posted Mar 8, 2013 7:36 UTC (Fri) by auc (subscriber, #45914) [Link] (8 responses)

Apple may be strong, but Google is dominant in mobile and desktop space, they own youtube. They can make a free video codec a reality.

Thank you, Google!

Posted Mar 8, 2013 12:10 UTC (Fri) by robert_s (subscriber, #42402) [Link] (7 responses)

Yet they haven't.

How long has WebM been around? What percentage of youtube videos are still not offered in WebM?

Thank you, Google!

Posted Mar 8, 2013 14:18 UTC (Fri) by drag (guest, #31333) [Link] (6 responses)

All of them?

I have yet to run:

'clive -f best http://www.youtube.com/watch?v=foo' and not get a WebM version of the video.

I think that one of the major reasons why you don't see more WebM by default is by the technical limitations of HTML5 versus Flash for the purposes of overlays and such things for adverts, although I could easily be mistaken.

Adverts

Posted Mar 8, 2013 19:13 UTC (Fri) by tialaramex (subscriber, #21167) [Link]

Yup, it's all about the adverts. If you embed a Youtube video then it doesn't come with adverts in most cases. HTML 5 browsers that implement WebM but not H.264 get sent WebM video and everything works.

Now try the same exact video, same browser, but found on Youtube directly, most likely you will get a message saying that you need Adobe Flash to play the video. We've seen that Youtube has a WebM version of the video, so why the message? Because Flash is used to display and control advertisements and Youtube is fuelled by advertising revenue.

Thank you, Google!

Posted Mar 8, 2013 19:30 UTC (Fri) by robert_s (subscriber, #42402) [Link] (4 responses)

That _is_ interesting. I didn't know the WebM versions were there at all.

Thank you, Google!

Posted Mar 11, 2013 8:06 UTC (Mon) by jezuch (subscriber, #52988) [Link] (3 responses)

> That _is_ interesting. I didn't know the WebM versions were there at all.

In youtube-dl it's -f 43 or -f 45 :) By default the H.264 version is considered "best quality" and going after best quality is the default. I change that whenever I can (not all videos are transcoded to WebM).

Thank you, Google!

Posted Mar 11, 2013 23:14 UTC (Mon) by lsl (subscriber, #86508) [Link] (2 responses)

Note that at least Fedora's package ships a config file (with --prefer-free-formats) that leads to downloading WebM files in what feels like 99 % of the time.

Thank you, Google!

Posted Mar 18, 2013 16:49 UTC (Mon) by wookey (guest, #5501) [Link] (1 responses)

This is interesting stuff. I find that somewhere around 40% of what I want to watch works (i.e is html5+ (presumably) webM). The rest I get a 'play flash' button that never gets past the 'wait' rotatey icon, due no doubt to gnash/lightspark incompatibilities). Clive doesn't work at all for me since some recentish (4 months ago?) UI changes (and niether does get_flash_video). youtube-dl does work (for both flash and html5 vids).

All this is very tiresome. I'd love to be able to tell youtube 'show me which need flash' or even 'don't show me any that need flash as I can't play them without a load of faff.' Some reliable way to get things to either 'work for everything' or 'only show stuff that will actually work' would be a huge boon. But my understanding of how all this fits together is too weak to do much more than take what I'm given.

I'd even be happy to spend some time developing this to fix things if I had any idea how to. Does removing gnash/lightspark help because now your browser stops claiming to be able to do flash?

Thank you, Google!

Posted Mar 18, 2013 17:35 UTC (Mon) by njwhite (guest, #51848) [Link]

> I'd even be happy to spend some time developing this to fix things if I had any idea how to. Does removing gnash/lightspark help because now your browser stops claiming to be able to do flash?

Nope. Instead you just get a "you need flash to play this video" message when you try to load the video page, so no improvement on your current setup. It is enormously annoying.

Though the experience is still not as bad as when using the official flash player, which inserts adverts all over the place - I don't know how people viewing things that way tolerate it.

VP8 and MPEG LA (WebM blog)

Posted Mar 8, 2013 8:33 UTC (Fri) by Seegras (guest, #20463) [Link]

I don't understand the issue.

The USPTO issued patents on mathematics illegally to MPEG-LA, and now somebody even takes these officially invalid patents into consideration?

Really, the only thing to do would be to take the USPTO to justice for its illegal doings.

VP8 and MPEG LA (WebM blog)

Posted Mar 8, 2013 12:04 UTC (Fri) by Company (guest, #57006) [Link] (8 responses)

So that means that VP8 is officially patented now and because the patent license is not GPL-Free (as it restricts what you can do with the source code), we'll need to drop it from Fedora now?

Or did I get that wrong? Help, I don't speak lawyer, someone explain the legal implications to me please.

VP8 and MPEG LA (WebM blog)

Posted Mar 8, 2013 13:26 UTC (Fri) by dowdle (subscriber, #659) [Link]

vp8 was always patented... they just promised to never seek royalties by granting free use of them to anyone. That is my paraphrase which is obviously not in the correct legal mumbo jumbo... but hopefully you get the point.

VP8 and MPEG LA (WebM blog)

Posted Mar 8, 2013 15:23 UTC (Fri) by lambda (subscriber, #40735) [Link] (6 responses)

No, I don't think that there is any admission that VP8 is covered by these patents, just that if so, there is a license to use it regardless.

Also, what makes you say that the license is GPL incompatible? As far as I know, the only requirement of the GPL is that if you have a patent license, you must grant the same rights to people downstream, and it sounds like this license does that. Do you have details of the agreement that indicates that it's not GPL compatible?

VP8 and MPEG LA (WebM blog)

Posted Mar 8, 2013 21:30 UTC (Fri) by Company (guest, #57006) [Link] (4 responses)

GPL allows turning a VP8 encoder into an H264 encoder. The patent license only covers VP8, no?

VP8 and MPEG LA (WebM blog)

Posted Mar 9, 2013 0:36 UTC (Sat) by xnox (guest, #63320) [Link] (1 responses)

If by writing code, you violate a patent, the license of your code does not prevent that, nor save/help you.

VP8 and MPEG LA (WebM blog)

Posted Mar 16, 2013 7:53 UTC (Sat) by drag (guest, #31333) [Link]

Yes. Patent law and copyright law are entirely different items. People lump them together and call them all 'IP', but it's not a accurate representation of how the law works.

Saying that you can take VP8 code and combine it with x264 to make a patent-free H.264 implementation (or whatever your approach is) is like saying: I paid my taxes this year so I can speed on the highway without fear of getting a ticket.

GPL and Patents

Posted Mar 9, 2013 2:00 UTC (Sat) by ARealLWN (guest, #88901) [Link]

Yes, the GPL itself doesn't restrict you from turning a VP8 encoder into an H264 encoder. Patent laws restrict your legal ability to do so. As I understand it the GPL only covers granting rights to implement the current software and patents applicable to it's current implementation, not all possible permutations or extensions of the released code. Therefore, I believe granting permissions to include VP9 codecs is an extension of patent grants not required by the gpl. I see no reason why these patent grants would be gpl incompatible. I have absolutely no formal legal training and do not claim this to be sound legal advice. If you are uncertain of any legal issues please contact a trained solicitor/lawyer.

VP8 and MPEG LA (WebM blog)

Posted Mar 9, 2013 3:00 UTC (Sat) by dlang (guest, #313) [Link]

well in that case you have to drop everything from Fedora, because the GPL allows you to convert any software into a H264 encoder.

This includes all the software that Redhat owns the patents on (or purchased a license that covers all GPL uses as well)

VP8 and MPEG LA (WebM blog)

Posted Mar 9, 2013 21:56 UTC (Sat) by tterribe (guest, #66972) [Link]

> No, I don't think that there is any admission that VP8 is covered by these
> patents, just that if so, there is a license to use it regardless.

There's actually an argument to be made that it confirms they're _not_ covered by these patents: http://xiphmont.livejournal.com/59893.html

VP8 and MPEG LA (WebM blog)

Posted Mar 10, 2013 1:56 UTC (Sun) by gdt (subscriber, #6284) [Link]

Another point to consider is that MPEG LA operates under the terms of a Business Review Letter from the USDOJ Antitrust Division.

In attempting to establish a patent pool for VP8 MPEG LA could be moving beyond the terms of that Letter. I am sure Google would argue to DOJ that MPEG LA was already operating beyond the limits of that Letter, thus by definition had breached the anti-trust laws in the view of DOJ, and thus should be prosecuted, and if the prosecution is successful, MPEG LA disbanded.

Of course should MPEG LA not seek a patent pool for VP8 and successors, then Google would not find it worthwhile to hand this "brief we have prepared" to DOJ or to pursue a private antitrust suit using the insider materials about MPEG LA which came to Google via the Motorola acquisition.

This scenario would explain the lack of a patent cross-licensing, because the issue under negotiation here isn't MPEG LA's patents, but the continuing existence of MPEG LA itself. Indirectly this would also turn MPEG LA from a revenue source for Apple, etc into a loss, since MPEG LA's legal fees would need to come from somewhere. Thus the pressure on MPEG LA to reach an agreement rather than to continue with FUD.

All of the above is, of course, speculation.